Successful planning and decision making in many technical and other industries depends on the expeditious and correct interpretation of complex information. For example, in the drug industry the data may have origins as diverse as high throughput screening experiments, clinical trials, patent information and research journals. In the petroleum industry the data may span seismic measurements, aerial surveys, laboratory data and economic forecasts. A system capable of providing unified access to disparate data sources and applications reduces the time spent finding, accessing, preparing, transforming and reformatting data, and allows professionals to focus on the interpretation and extraction of knowledge for planning and decision making.
However, one complication with providing this type of unified access is that the data inevitably spans several disciplines, with an attendant probability of misinterpretation. Extensive knowledge of multiple domains is required if misuse is to be avoided.
Therefore, there is still an unsatisfied need for an information management system that clarifies the generation, use, and purpose of the data. The information management system can capture knowledge about the genesis and history of the data, how analyses are done, how decisions are made, and what the outcomes are. This “corporate memory” forms the basis for the analysis required to make better technical and business decisions.
Several attempts have been made to access information based on annotations. Illustrative attempts are described in the following references:    U.S. Pat. No. 5,404,295 to Katz et al.    U.S. Pat. No. 5,600,775 to King et al.    U.S. Pat. No. 5,832,474 to Lopresti et al.    U.S. Pat. No. 5,548,739 to Yung.
For example, U.S. Pat. No. 5,404,295 describes a method and apparatus for computer retrieval of database material. Annotations are provided for selected database subdivisions and are converted to a structured form and stored in that form along with connections to corresponding subdivisions. Searching for relevant subdivisions involves entering a query in natural language or structured form, converting natural language queries to structured form, matching the structured form query against stored annotations, and retrieving database subdivisions connected to matched annotations.
However, the teaching of this patent is limited to a system with the capability to search the annotations to locate the database material. The system does not have the capability to search the stored information based on both the annotations and database material, or to search on database material to retrieve the annotations. As a result, the system is not suitable for directly locating a subset of data where the filter has predicates on both the annotations and database material. Rather, it will locate all database material that corresponds to the annotation predicates and it would require a second step to filter this subdivision and to apply the data predicates.